Our overall objective is to conduct more directed studies of main effects of genes and gene-gene and gene environment interactions within selected metabolic pathways of interest in the etiology of colorectal adenomas, accepted precursors of colorectal cancer (CRC). In the parent study for this application, we have DNA and risk factor data on 1,610 subjects (778 cases and 832 controls). We have the following specific aims: 1) Study selected genes and environmental factors relevant to a folate-related pathway. Specifically, we propose to study MTHFR, MTRR, ADH3, which may affect folate metabolism, and XRCC1, XRCC3, OGG1, and MGMT, which may be involved in DNA repair of damage associated with folate metabolism. In addition, we have data already in hand on RBC folates, homocysteine, and DNA methylation that will be incorporated into statistical analyses of these data where appropriate. 2) Study genes that may be involved in the metabolism or transport of bile acids. Specifically, we propose to study SLC10A2, the gene encoding the ilieal sodium-dependent bile acid transporter (ISBT), along with the VDR and EPHX1 genes. We will also conduct analyses to investigate if the main effects of these genes appear to be modified by selected environmental factors that have been hypothesized to affect risk of colorectal adenomas or cancer via an effect on bile acids. 3) To characterize systematically in vitro all non-synonymous SNPs in selected genes of epidemiologic interest, such as EPHXl (alias mEH) and MTHFR, by site-directed mutagenesis of the cloned enzyme, overexpression and appropriate assays, and to characterize systematically in vitro all SNPs in the 5' and 3' UTRs (untranslated region) of selected genes of epidemiologic interest in appropriate reporter constructs, e.g. expressing modified firefly luciferase. Our rationale for selecting alleles to be characterized is described in the Study Design and Methods section; 4) As a statistical method to better analyze complex metabolic pathways is further developed by Drs. Thomas and Cortessis, we will apply this method to data generated by this study.